Drunk drivers will almost always get home safely each night. And if that happens enough successive times, the human brain “learns” a very bad thing, that it is okay to be driving drunk. Until tragedy strikes and it’s not. The Barack Obama foreign policy had been famously summarized as “Don’t do stupid s**t,” which is opposite of drunk driving, but that drunk driving has similar risks to the foreign policy of Donald Trump.
Tragedies such as the downing of a Ukrainian passenger jet over Iran last week frequently follow as unintended results of high-risk, ill-advised actions, such as Trump’s termination of an (albeit nefarious) top Iranian official. We scramble to assign fault and blame, yet we rarely see the issue as one of probability and the inevitability of what happens after we do the “stupid stuff.” So, let’s go down that road, first with a history lesson about risk and tragedy, and then some math.
“Take a trip with me to 1913…”
I have been to the old Upper Peninsula Michigan mining town of Calumet many times, but until last September I had never stopped at the corner of Elm and 7th Street to see the somber memorial installed there in recent years by the National Park Service. The arch and cornerstone shown above are all that remain from a building called Italian Hall. This was the site of a tragic event in 1913 called either the “Italian Hall disaster” or the “Calumet Massacre of 1913” depending on your union viewpoint at the time. My father was born two years after this event and only 15 miles away, and he had family living not far from to this site.
On Christmas Eve of 1913, a party was being held in the top floor of this building for the children of striking copper miners. After a false cry of false “Fire!” and the resultant panic, 73 people, mostly the children, were killed in a stampede trying to escape the building. Two years ago, I reviewed a new book about this incident by Daniel Wolff entitled Grown-Up Anger: The Connected Mysteries of Bob Dylan, Woody Guthrie, and the Calumet Massacre of 1913 (HarperCollins, 2017). Folksinger Woody Guthrie wrote a song about the tragedy called “The 1913 Massacre,” which his son Arlo later recorded.
Daniel Wolff recounts the complexity of this tragedy well. Tensions were high between striking union workers and their bosses. Strike breakers were in town, and they caused trouble daily. The workers were mostly immigrants or their sons, very poor and poorly paid. They worked in unsafe conditions in mile-deep copper mines, with most of their families living in bad “company town” housing.
My great-grandfather, an immigrant from far northern Norway himself, had been a miner in the next town to the south of Calumet, and his daughter, my grandmother, called one such company town neighborhood near where she grew up “Helltown.” Italian Hall was unsafe and overcrowded, with one narrow stairway and likely a door that opened inward. In the panic of the “Fire!” shout (maybe/maybe not from the strikebreakers?), children died by the dozens on the stair.
It was the fog of a distant war, not between countries, but between two different social systems, the open economic war in that period between “Labor” and “Capital.” For both sides, tempting fate through risky actions and “trigger events” was on the daily menu. There was likely “no one cause” for the Italian Hall disaster. The Fates, or “God,” or “stochastic processes” had plenty of bad outcomes to choose from. No one was to blame, and everyone was to blame.
All of these factors raised the “probability temperature” of tragedy, ready for a “match” to be lit. You can call this stochastic tragedy, akin to stochastic terrorism, which I have written about in the past.
“Wait for it…wait for it…boom!”
(If you want to skip the math and go right to the foreign affairs stuff, here is your chance to jump to the next part.)
Countless natural and man-made events we encounter every day are called Poisson processes (pronounced “pwa-sahn”), named after French mathematician Siméon Denis Poisson (1781–1840). These are low-frequency events that seem to happen randomly, and yet with a predictable pattern to that randomness that emerges over time. Natural Poisson events include the replication of a skin cell among one small patch of skin, or the time until the collapse of a sandpile as more sand is pouring in. It could be the “firing” of a set of your brain neurons triggering a body movement, or even, as Poisson himself found, the rate at which horses kick cavalry soldiers. Man-made events that often follow Poisson’s predicted pattern include calls coming into a telephone hotline and, sadly, even the occurrence of mass shootings.
For infrequent events, that Poisson “hit” may or may not happen in the very next tiny instant of time, or the next, but eventually the “wait for it, wait for it…” ends and, boom!, the next call comes into the telephone support line. For any given kind of Poisson event, for instance the replication of a skin cell within a small patch of skin under a microscope, an average rate of “hits” will emerge over time, designated by the Greek letter lambda ( λ ) in the examples below. When you graph it out, the curve looks like a classic “normal” distribution that gets increasingly “squashed” on the left as the probability of a “hit” within the next period of time drops. And, as we will see, that “squashed” shape is important.
Drunk drivers meet Monsieur Poisson
How many sober drivers will die in your state in a traffic accident tomorrow during daylight hours in dry conditions on non-busy roads? You can think of this example as like the yellow dotted line above. There are good odds that zero, or perhaps one, may die, out of millions of drivers on the road. I call these “lottery odds.” You don’t want to “win” this lottery, but thankfully, you probably won’t because the Poisson probability is tiny.
It’s a different story, however, late on a weekend night in bad road conditions. Most drunk drivers will still get home safe, as I noted, but the probabilities of an accident, and perhaps dying, have risen by several “orders of magnitude” (orders of ten), which is a lot. Note that the probability of zero events happening drops as the expected number of “hits” rises. A weekend night without a drunk driving death in Florida is a rare event.
Smoking tobacco produces a similar pattern. Contrary to popular opinion, most heavy smokers die of something other than lung cancer, but frequent tobacco use moves you from “lottery” odds to “drunk driver” odds, or even higher. Are you a betting person?
Diplomacy and Poisson probabilities
To interpret that graph above in social and international terms, you cannot eliminate all tragedies like terrorist incidents or gun violence. However, you can usually push the probability of zero events happening within the next week as high as possible, on the orange curve above. And you can certainly get the probability of multiple incidents down the scale from the blue curve to a near-zero chance. You do this by learning what the high-risk “trigger” events are (i.e., the “stupid s**t”) and at least postponing them as long as possible. With bad Poisson events, time is your friend.
Every evening that a tipsy bar customer takes an Uber home or that an illegal gun is not sold on the street literally changes the probability curve on something bad being triggered that day. When it comes to international relations, every day that Secretaries of State are talking rather than authorizing violent response lowers the probability of a tragic incident, or worse, happening tomorrow. But Donald Trump and Mike Pompeo like to play with matches.
Other people’s money and moral hazard
Risky behavior inevitably gets worse when people are playing the game with “other people’s money.” Insurance actuaries call this a “moral hazard.” Some clown who has fire insurance on his building will burn it down in an attempt to collect. A co-worker of mine once decided to test the performance limits of his rental car in a snowstorm, a risk he never would have taken with his own vehicle. The test did not turn out well. That’s moral hazard.
Donald Trump has always done his riskiest deals with “other people’s money,” and that, too, has frequently turned out badly. When it is someone else’s kids on the battle line rather than your own, and someone else’s money paying for the bombs and the cleanup afterward, then the Poisson probabilities of both bellicose action and tragic outcomes climb precipitously.
Somebody else shot down the Ukrainian airliner, but Donald Trump was lighting the matches. It is time to take the matches away from the children.
- Stochastic terrorism part 1 – The sand pile effect
- Stochastic terrorism part 2 – the mass shooting lottery
- Worth a read: “Grown-up Anger” by Daniel Wolff